Author: Sylvia Kaufmann Title: Factor augmented VAR revisited - Exploiting sparsity to include meaningful factors Abstract: We combine the factor augmented VAR framework with recently developed estimation and identification procedures for sparse dynamic factor models. Induced sparsity in the factor loading matrix helps identifying the factor basis, without imposing additional rotational identification restrictions, and allows us to estimate an unrestricted, reduced-form factor augmented VAR for a large panel of U.S. economic data. Column-wise, the non-zero factor loadings provide a meaningful interpretation of factors. We combine two structural identification methods to further analyze the model estimate. Responses of factors and specific variables to a monetary policy, a term premium and a productivity shock are sensible and consistent with previous studies. Comparing ours to other parametric and non-parametric factor estimates uncovers some advantages of parametric sparse factor estimation for large economic datasets.